Web and Social Network Analytics

Week 5: Ethics in Social Network Analytics

Dr. Zexun Chen
Zexun.Chen@ed.ac.uk

Feb-2026

Table of Contents

Welcome Back!

Last Week:

  • Sentiment analysis for customer reviews
  • Association rules and frequent itemsets
  • Recommendation systems

This Week: Ethics in Analytics

  • The responsibility that comes with data
  • GDPR and privacy regulations
  • Fairness in algorithmic systems

Learning Objectives

By the end of this lecture, you will be able to:

  1. Explain why ethics matters in web and social network analytics
  2. Identify key GDPR requirements for data collection
  3. Recognise sources of algorithmic bias in AI systems
  4. Apply Privacy by Design principles to ShopSocial

Privacy & Data Protection

Why Ethics Matters

The Stakes Are High

Every click, scroll, and purchase generates data. How we collect, store, and use this data has real consequences for individuals and society.

  • Trust: Users who feel surveilled abandon platforms
  • Reputation: Data breaches and misuse destroy brands
  • Legal: GDPR fines can reach €20 million or 4% of global revenue
  • Societal: Unchecked tracking enables manipulation and discrimination

Case Study: Cambridge Analytica (2018)

What happened:

  • Facebook quiz app harvested data of 87 million users
  • Used for political advertising without proper consent
  • Influenced Brexit and US elections

Consequences:

  • Facebook fined $5 billion (US)
  • Cambridge Analytica dissolved
  • GDPR enforcement accelerated

Lessons for ShopSocial:

  1. Consent must be explicit
  2. Third-party data sharing is risky
  3. Users expect transparency
  4. Build ethically from the start

đź“Š Quick Poll: Privacy Trade-offs

Wooclap Question

Go to wooclap.com and enter code: KMIAHE

What data would you share for a 10% discount on ShopSocial?

A. Purchase history only

B. Location data

C. Browsing behavior

D. None - privacy is more important

GDPR: Key Requirements

Privacy by Design: 7 Principles

  1. Proactive not reactive - prevent problems
  2. Privacy as default - most protective settings
  3. Privacy embedded - integral, not an add-on
  4. Full functionality - positive-sum approach
  5. End-to-end security - protect data lifecycle
  6. Visibility & transparency - open practices
  7. Respect for user privacy - user-centric

Applied to ShopSocial:

  • Only track essential metrics
  • Anonymise IP addresses
  • Provide clear opt-out mechanisms
  • Regular data deletion schedules

Fairness & AI Regulation

Fairness in AI: Real Cases

Amazon’s AI Hiring Tool:

Shut down after discovering gender bias - the system penalized female candidates.

ProPublica Investigation:

Recidivism risk scoring was biased against African Americans.

Algorithmic Bias: Sources

\[Data + Model = Prediction\]

Common Bias Sources:

Type Example
Selection Bias Training data doesn’t represent population
Feedback Loops Recommendations reinforce existing patterns
Omitted Variables Missing attributes that affect outcomes
Societal Bias Human biases embedded in historical data

Key Insight: Simply ignoring sensitive attributes (gender, race) does NOT remove bias - it’s already embedded in other features.

🔍 Hands-On: Ethical Audit

Activity (3 minutes)

You’re auditing ShopSocial’s analytics. Assess each feature:

Feature Ethical Concern Acceptable?
Personalized pricing Price discrimination
“Customers like you bought” Filter bubble
Abandoned cart emails Manipulation
Sharing data with advertisers Privacy violation

Discussion: Which features would you approve? What safeguards would you add?

EU AI Act: Risk-Based Regulation

Note: penalties: Up to €35M or 7% of global annual turnover.

GDPR vs EU AI Act

Aspect GDPR EU AI Act
Focus Personal data & privacy AI systems & risks
Scope Data collection/processing AI model development
Key Concern Consent & data rights Fairness & safety

For ShopSocial: Must comply with BOTH - protect customer data AND ensure fair AI recommendations.

Summary

Key Takeaways

Privacy:

  • GDPR protects user rights
  • Consent must be explicit
  • Privacy by Design principles

Fairness:

  • Bias enters through data AND algorithms
  • Ignoring attributes doesn’t fix bias

Regulations:

  • GDPR: data protection
  • EU AI Act: AI risk classification
  • Significant penalties for non-compliance

Practice:

  • Ask: Do we need this data?
  • Transparency builds trust

Course Wrap-Up

What You’ve Learned:

Week Topic
1 Web Analytics Fundamentals
2 Web Scraping & APIs
3 Social Network Analysis
4 Unsupervised Learning
5 Ethics & Regulations

Next Steps:

  • Apply these techniques to your ShopSocial coursework
  • Consider ethics at every stage of your analysis

Questions?

Thank you!

Dr. Zexun Chen

Zexun.Chen@ed.ac.uk

Office Hours: By appointment

Good luck with your coursework!